Files
Fortran-docker-mvc/flibs-0.9/flibs/doc/datastructures/mem_pool.html
Tommy Parnell a7037b9e92 init
2017-02-20 16:06:45 -05:00

111 lines
3.8 KiB
HTML

<html><head>
<title>mem_pool - flibs </title>
</head>
<! -- Generated from file 'mem_pool.man' by tcllib/doctools with format 'html'
-->
<! -- Copyright &copy; 2006 Arjen Markus &lt;arjenmarkus@sourceforge.net&gt;
-->
<! -- CVS: $Id: mem_pool.html,v 1.1 2008/06/13 10:33:28 relaxmike Exp $ mem_pool.n
-->
<body>
<h1> mem_pool(n) 1.0 &quot;flibs&quot;</h1>
<h2><a name="name">NAME</a></h2>
<p>
<p> mem_pool - Implement a straightforward memory pool
<h2><a name="table_of_contents">TABLE OF CONTENTS</a></h2>
<p>&nbsp;&nbsp;&nbsp;&nbsp;<a href="#table_of_contents">TABLE OF CONTENTS</a><br>
&nbsp;&nbsp;&nbsp;&nbsp;<a href="#synopsis">SYNOPSIS</a><br>
&nbsp;&nbsp;&nbsp;&nbsp;<a href="#description">DESCRIPTION</a><br>
&nbsp;&nbsp;&nbsp;&nbsp;<a href="#data_types_and_routines">DATA TYPES AND ROUTINES</a><br>
&nbsp;&nbsp;&nbsp;&nbsp;<a href="#implementation_notes">IMPLEMENTATION NOTES</a><br>
&nbsp;&nbsp;&nbsp;&nbsp;<a href="#copyright">COPYRIGHT</a><br>
<h2><a name="synopsis">SYNOPSIS</a></h2>
<p>
<table border=1 width=100% cellspacing=0 cellpadding=0><tr bgcolor=lightyellow><td bgcolor=lightyellow><table 0 width=100% cellspacing=0 cellpadding=0><tr valign=top ><td ><a href="#1"><b class='cmd'>call pool_acquire( pdata )</b> </a></td></tr>
<tr valign=top ><td ><a href="#2"><b class='cmd'>call pool_release( pdata )</b> </a></td></tr>
</table></td></tr></table>
<h2><a name="description">DESCRIPTION</a></h2>
<p>
The <em>mem_pool.f90</em> source file defines a set of subroutines
that allow you to acquire and release memory of a particular derived
type, thereby reducing the number of allocations and deallocations.
Such a memory pool is useful for instance when you need temporary memory
of fixed size.
<h2><a name="data_types_and_routines">DATA TYPES AND ROUTINES</a></h2>
<p>
The source code expects a data type, POOL_DATA, that contains an integer
field &quot;pool_index&quot; for private use by the subroutines. All other fields
can be used by the application itself:
<p><table><tr><td bgcolor=black>&nbsp;</td><td><pre class='sample'>
module MYDATA_POOL
type POOLDATA
integer :: pool_index ! For private use by pool_acquire/pool_release
real, dimension(100) :: work ! The actual work space
end type
include &quot;mem_pool.f90&quot;
end module MYDATA_POOL
</pre></td></tr></table></p>
The code defines the following routines:
<dl>
<dt><a name="1"><b class='cmd'>call pool_acquire( pdata )</b> </a><dd>
Get a pointer to available data in the memory pool. The memory pool is
expanded automatically, by allocating an array of 100 items. If there
is an error (no more memory can be allocated), the pointer will not be
associated.
<br><br>
<dl>
<dt>type(pool_data), pointer <i class='arg'>pdata</i><dd>
The pointer variable that will be associated with valid memory
</dl>
<br><br>
<dt><a name="2"><b class='cmd'>call pool_release( pdata )</b> </a><dd>
Release the data pointed to by pdata to the memory pool. This data will
become available again for further acquiring.
<br><br>
<dl>
<dt>type(pool_data), pointer <i class='arg'>pdata</i><dd>
The pointer variable pointing to data to be released into the pool
</dl>
</dl>
(Note: two more subroutines are envisioned: setting two parameters that
control the allocation and deallocation of memory, and a routine to
print statistical information. These have not been implemented yet.)
<h2><a name="implementation_notes">IMPLEMENTATION NOTES</a></h2>
<p>
The subroutines do not change the fields of the POOL_DATA structure
(except for pool_index). This means that allocatable (pointer)
components in this structure are not influenced. You could use this to
store dynamically sized arrays in the memory pool. It is especially
useful if all such arrays have the same size.
<h2><a name="copyright">COPYRIGHT</a></h2>
<p>
Copyright &copy; 2006 Arjen Markus &lt;arjenmarkus@sourceforge.net&gt;<br>
</body></html>